Conventionally, in order to generate a cartoon-like (comic-like) image by three-dimensional image processing, a process of drawing an outline (contour) of a model formed of polygons has been proposed. For example, in image processing apparatuses described in Patent Documents (Japanese Patent Laid-Open Publications Nos. 2000-251094, 2001-126086 and 2001-084404), a model indicating a character or the like (basic model) and an enlarged model for drawing an outline are used. The enlarged model is obtained by enlarging the basic model and is arranged to encompass the basic model. Also, the normal vectors of polygons constituting the enlarged model are reversed with respect to the normal vectors of the basic model. Specifically, the normal vectors of the basic model are set to be directed outward while the normal vectors of the enlarged model are set to be directed inward. Thereby, when the basic model and the enlarged model are drawn by three-dimensional image processing using hidden surface removal or the like, only a portion(s) of the enlarged model which is located behind the basic model as viewed from a viewpoint and outside the edge of the basic model is drawn. Therefore, the enlarged model may be, for example, filled with black or the like, thereby making it possible to generate an image with an outline attached to the outside of the basic model.
In the image processing apparatuses of Patent Documents above, two models, i.e., the basic model and the enlarged model, are required to draw one character, so that additional processes are required as compared to when one model is used for one character. Therefore, the larger the number of polygons constituting one model, the more the processing load. Also, the larger the number of models to be drawn, the more the processing load.
Also, in the image processing apparatuses of Patent Documents above, a portion of the enlarged model outside the basic model becomes an outline. Therefore, the thickness of the outline in the generated image varies depending on the distance from the viewpoint to the model. Therefore, if the distance from a camera to the model is large, the outline may be too thin to be visually recognized. Conversely, if the distance is excessively small, the outline may be excessively thick.